The Expression of Evidentiality in French-English Bilingual Discourse R U T H K I N G Dept
Total Page:16
File Type:pdf, Size:1020Kb
Language in Society 28, 355–365. Printed in the United States of America The expression of evidentiality in French-English bilingual discourse R U T H K I N G Dept. of Languages, Literatures & Linguistics, York University Toronto, Ontario M3J 1P3, Canada [email protected] T E R R Y N A D A S D I Dept. of Modern Languages & Cultural Studies, University of Alberta Edmonton, Alberta T6G 2E1, Canada [email protected] A B S T R A C T This study, drawing on data from a large sociolinguistic interview corpus for three Acadian communities of Atlantic Canada, concerns codeswitches involving verbs of opinion or belief (e.g. guess, think, imagine, believe) in French-English bilingual discourse. The codeswitch itself serves to under- score the speaker’s stance as to the truth of the proposition – and, in some cases, to indicate a degree of uncertainty not nuanced by corresponding French language forms. Variation in usage is related to intensity of lan- guage contact at the levels of the community and of the individual. (Code- switching, discourse analysis, evidentiality, quantitative sociolinguistics, Canada, Acadian, French)* Most recent research on codeswitching has been conducted from one of two per- spectives: grammatical or interactional. The search for grammatical constraints on intrasentential codeswitching exemplifies the grammatical perspective (e.g. Poplack 1980, DiSciullo et al. 1986, Poplack et al. 1989, Myers-Scotton 1993, Belazi et al. 1994, Mahootian & Santorini 1996), while the study of the social meaning of particular codeswitches exemplifies the interactional perspective (e.g. Gumperz 1982, Heller 1982, 1988, 1994). The present article concentrates in- stead on the role of codeswitching in the organization of discourse; specifically, it is concerned with how codeswitching figures in the expression of evidentiality in French-English bilingual discourse. Following Auer 1995, we argue that code- switching can be analyzed at the level of discourse, relatively independently of any grammatical properties of codeswitching or of the immediate social context in which it is embedded. We do turn, however, to consideration of the sociolin- guistic situation in order to explain inter- and intra-community variation. © 1999 Cambridge University Press 0047-4045099 $9.50 355 R U T H K I N G A N D T E R R Y N A D A S D I T H E C O R P U S The present study is concerned with language use in three Acadian communities of Atlantic Canada, two in the province of Prince Edward Island (PEI) (Abram- Village and Saint-Louis), and one in Newfoundland (L’Anse-à-Canards). All are small fishing villages with fewer than five hundred residents. Although the gram- mars of these three varieties are the same for those linguistic features relevant for our study, the relative prestige of French differs in the three communities, as does the degree of contact with English. L’Anse-à-Canards is situated in an isolated part of Newfoundland where there was little contact with English before World War II.1 Until quite recently, educa- tion in French has been negligible. Elsewhere (King 1989, King & Nadasdi 1997) we have shown that Newfoundland French is one of the most conservative Aca- dian varieties as regards influence from English or from other varieties of French, including the standard. The two PEI varieties under discussion have been in closer contact with English, for a longer period, but there are striking differences be- tween the two communities. Abram-Village is situated in a small enclave, the Evangéline region, where French is the majority language locally, although it is in a minority position in the province as a whole. There is strong institutional support for French, and thus our sample includes speakers with some control of the standard as well as of the local variety. Saint-Louis, by contrast, is surrounded by English-speaking villages, and there is little institutional support for French. Not surprisingly, lack of transmission of the language to the young is a serious problem faced by the community. Speakers of Saint-Louis French have had more exposure to English, but less exposure to Standard French, than their counter- parts in Abram-Village. A sociolinguistic interview corpus for twenty-four Abram-Village residents and twenty Saint-Louis residents, comprising a total of just over 800,000 words, provides the main data for our study; a subsample of our L’Anse-à-Canards cor- pus, comprising interview data for eight residents and consisting of just over 100,000 words, was used for comparative purposes. S W I T C H E D F O R M S The data in exx. 1–3 come from our interview corpus.2 In all cases, the matrix language is French: (1) I guess qu'on est pas mal tout pareil. (19.2A.255, Abram-Village) ‘I guess that we are just about all equal.’ (2) I think j'ai plus peur des chenilles qu'une serpent.3 (30.2A.47, Saint-Louis) ‘I think (that) I’m more afraid of caterpillars than a snake.’ (3) I imagine qu'il y en a qui l'avont encore. (29.1B.269, Saint-Louis) ‘I imagine that there are some who still have it.’ In the PEI corpora, one finds codeswitches like I guess, I imagine, I think, I bet, and I'm sure with French that-clause complements. Such tokens all involve a 1sg. 356 Language in Society 28:3 (1999) E V I D E N T I A L I T Y I N F R E N C H - E N G L I S H B I L I N G U A L D I S C O U R S E pronoun; i.e., one does not find examples like 4, with a 1pl. pronoun, or 5, with a lexical NP as subject: (4) *We guess qu’on est pas mal tout pareil. `We guess that we are just about all equal.' (5) *Marie doesn’t know quoi ce qu’a arrivé. `Marie doesn't know what happened.' In addition to the matrix clause use of English codeswitches, there is also widespread use of emblematic, or tag, switching, involving I think, I guess, I imagine, etc. at the ªedgesº of sentences (as in 6±8). We also find sentential codeswitching (9±10): (6) Ils avont pas mal de la misère, I guess. (01.1B.407, Abram-Village) `They are having a hard time, I guess.' (7) J’étions une quarantaine, I suppose, une quarantaine. (30.1A.108, Saint-Louis) `There were about forty of us, I suppose, about forty.' (8) C’est sept ou huit heures, je sais pas, huit heures, I imagine. (33.1B.810, Saint-Louis) `It's seven or eight hours, I don't know, eight hours, I imagine.' (9) A: Les Français alentour d’icitte s’accordont bien? `(Do) the French around here get along well?' B: Bien, I guess so. `Well, I guess so.' A: Ils travaillont ti ensemble? `Do they work together?' B: Oui, oui. (27.1B.208, Saint-Louis) `Yes, yes.' (10) A: Les traditions de la communauté comme la râpure puis les fricots puis toute ça, c’est ti de quoi qui va rester avec les jeunes? `Community traditions like râpure and fricot and all that, is that something that is going to stay with young people?' B: Ah oui! Je crois quasiment, je crois quasiment que oui. I think so. Je sais pas. (19.2B.273, Abram-Village) `Oh yes! I believe pretty much so, I believe pretty much so, yes. I think so. I don't know (for sure).' While our primary focus will be on intrasentential and emblematic codeswitches, it is worth noting that there are no striking differences in verb choice in single- clause utterances. English verbs employed in codeswitches are the following, in order of fre- quency: guess, think, don’t know, don’t think, imagine, believe, suppose, be sure, bet, and can’t see. We do not find other high-frequency English verbs, such as say, tell, ask, remember, show, or explain ± and this, we argue, is not accidental. T H E E X P R E S S I O N O F E V I D E N T I A L I T Y Regardless of syntactic position, we find that the choice of verb is semantically constrained: Codeswitches occur with a particular class of evidentials, namely verbs of opinion or belief. No such codeswitches are found with other classes of Language in Society 28:3 (1999) 357 R U T H K I N G A N D T E R R Y N A D A S D I verbs which take that-clause complements in English (cf. Partee 1973), such as verbs of communication (say, tell, explain), verbs of inference ( prove, show, discover), or emotives (be sad, be glad, hate). Thus, we do not find examples such as I said que or I showed que. Close analysis of the surrounding text leads us to suggest that, in many cases, the codeswitches serve to mitigate the speaker's relationship to the proposition ex- pressed in the embedded clause. In 11, the local interviewer asks the informant, a middle-aged Saint-Louis male, about his father's seeing the ghost of his first wife: (11) A: Ç’a ti ienque arrivé une fois ou – ? `Did that happen only once or ± ?' B: Bien ... as far as I know, oui. `Well ...as far as I know, yes.' A: Mmhmm. Il était ti marié là dans ce temps là ... à sa deuxième femme? `Mmhmm. Was he married then ... to his second wife?' B: I guess qu’il était marié avec la deuxième femme.